csvread 3.0.1 exponentially slower than v 3.0.0

[David Bateman]

gOS wrote:
> I've not sat through all the way, but on 2 seperate windows machines, one
> with a pentium 4 and one with a much older processor, I get similar results.
>
> I can confirm that in 3.0.0 the script took about 5-10 seconds to complete.
>
> In 3.0.1 I've sat for 10 minutes waiting for csvread to finish and it
> hasn't. Ctrl+c causes Octave to exit without warning and does not return
> control to the prompt.
>
> I'm reading in 3mb of text. I've attached said file.
>
> Command that hangs indefinately:
>
> data = csvread('myFile.dat');
>
> Help would be appreciated. I've confirmed that rolling back to 3.0.0 fixes
> the problem. I'm using the most up to date version from the Octave Forge
> site. Can anyone confirm that this is an Octave problem and not specific to
> windows? Can anyone one fix it? Thank you.
>
> I couldn't get a bug report to generate after calling csvread because it
> never finished. Here's my bug report for just starting up Octave though.
> Maybe it will help. :-/
>   
The change between the two versions was essentially to change the
C-string handling to C++ string handling. It also more correctly
duplicates the Matlab interface. This was done for this codes inclusion
in Octave itself.

Now the slow up appears impressive, and I'd like to see this situation
improved. However looking at the new dlmread.cc file used by csvread.m,
I can't see any obvious reason for the slow-up. Unfortuantely I don't
really have the time to look at this in the near future, so if the old
dlmread.cc was fine with your code I'd suggest getting it from

http://octave.svn.sourceforge.net/viewvc/*checkout*/octave/trunk/octave-forge/main/io/src/dlmread.cc?revision=4585

and using it instead. Alternatively, you might look at the new
dlmread.cc function and see how you might fix the slow-up issues.

D.

-- 
David Bateman                                David.Bateman at motorola.com
Motorola Labs - Paris                        +33 1 69 35 48 04 (Ph) 
Parc Les Algorithmes, Commune de St Aubin    +33 6 72 01 06 33 (Mob) 
91193 Gif-Sur-Yvette FRANCE                  +33 1 69 35 77 01 (Fax) 

The information contained in this communication has been classified as: 

[x] General Business Information 
[ ] Motorola Internal Use Only 
[ ] Motorola Confidential Proprietary